The most prevalent primary brain tumor, glioblastoma (GBM), ranks among the most lethal of human cancers. Like the normal brain, GBMs contain cellular hierarchies with self-renewing, multi-lineage cells at the apex. These brain tumor initiating cells (BTICs), or GBM stem cells, display therapeutic resistance, promote tumor angiogenesis, and invade into normal tissues. BTICs are not uniformly distributed, but rather concentrated in specific regions, or niches, including around blood vessels (the perivascular niche) and in perinecrotic (hypoxic) regions. These niches provide essential cues that maintain stem-like tumor cells through several pathways that appear specific for each location. In preliminary studies, we find that biopsies of specific GBM regions contain tumor cells with divergent transcriptional profiles and precursor markers. Cellular differentiation states reflect the cumulative regulation of chromatin, so we examined the expression and activity of two core epigenetic regulatory nodes: polycomb repressive complexes 1 and 2 (PRC1/2). Vascular tumor regions were associated with PRC2 activity, whereas hypoxic regions were associated with measures of PRC1. Collectively, these results suggest that stem-like cancer cells residing in different locations may be regulated by different epigenetic programs. Enhancer of zeste homolog 2 (EZH2) is a histone-lysine N-methyltransferase enzyme that serves as the enzymatic component of PRC2. B lymphoma Mo-MLV insertion region 1 homolog (BMI1) is an essential component of PRC1. Both EZH2 and BMI1 have been linked to normal and neoplastic stem cells, as well as therapeutic resistance. Clinical relevance for BMI1 and EZH2 is supported by the development of targeted therapies against each molecule. We find that mesenchymal BTICs display preferential activity and dependence on BMI1, whereas proneural BTICs display preferential activity and dependence on EZH2. As GBMs contain a mixture of proneural and mesenchymal tumor cells, we hypothesized that combined targeting of BMI1 and EZH2 will offer improved tumor control with acceptable toxicity. Further, we have identified a novel molecular regulatory mechanism of BMI1 linked to the hypoxic microenvironment that may inform patient selection for precision (personalized) medicine application of epigenetic targeting. Finally, additional studies will determine the potential of this therapeutic regimen to synergize with conventional therapies against glioblastoma. Collectively, these studies will investigate novel molecular regulation by the tumor microenvironment of neoplastic stem cells that can be directly translated into clinical trials.